Steven Den Beste presents evidence that polls are being jiggered to hep Kerry

USS Clueless ^ | October 12, 2004 | Steven Den Beste

[quidnunc]

Looking at the "RealClear Politics" plot of the presidential polls, I see two long term trendlines, punctuated by a hell of a lot of what I would refer to as "experimental error":

I don't believe that public opinion has been changing as much as these polls seem to suggest. The variation we see up through July looks like what engineers call "sample aliasing" or "jitter". Note that it falls well within the oft-claimed ±4 points of error. This is typical for data taken in noisy sampling environments; I've seen this kind of thing many times.

August and September are different. I've seen that kind of thing, too.

In my opinion, the polls were being deliberately gimmicked, in hopes of helping Kerry. In early August it looks as if there was an attempt to engineer a "post-convention bounce", but it failed and was abandoned after about two weeks. But I'm not absolutely certain about that.

The data for September, however, is clearly an anomaly. The data is much too consistent. Compare the amount of jitter present before September to the data during that month. There's no period before that of comparable length where the data was so stable.

The September data is also drastically outside of previous trends, with distinct stairsteps both at the beginning and at the end. And the data before the anomaly and after it for both Kerry and Bush matches the long term trendlines.

If I saw something like that in scientific or engineering data, I'd be asking a lot of very tough questions. My first suspicion would be that the test equipment was broken, but in the case of opinion polls there is no such thing. My second suspicion would be fraud.

-snip-

[quidnunc]

You need to go to the source document in order to see the graphic which shows what the author is talking about.

[Read2Know]

The polls conducted by politically biased organizations must be suspect. The clear motivator for skewing the polls is to save the party, not necessarily Kerry. If the base becomes too discouraged, they won't show up on election day. If the rank and file aren't there to vote, their people will be knocked out of office.

[Stellar Dendrite]

The media is such a propaganda machine, I wouldn't doubt that they were aware of these planned challenges for Nov. 2nd.
This is why they keep it as close as it is....it will put in public mind that the race was 'close' going in.
This way, the Bolsheviks can once again challenge the results in court claiming 'disenfranchisement' or whatever they feel like making up at the time. In the public's mind, it was already close to begin with....so the average American will accept this 'challenge' (remember, the slow boiling of the frog began in 2000).

[onyx eyes]

The polls are absolutely being "jiggered" to help kerry. Back in the day, when the Dems were running their primarys, Dean hee-hawed too loud, and all of a sudden kerry was there. What I noticed and being statistically- minded -- was that in each poll during the spring of 2004; in each the contested states for the primarys, kerry was always leading and nearly always had a poll number of 47%. It got freaky after awhile.

[xzins]

Interesting.

But it doesn't answer the question, "Why would the pollsters, then, be honest at any time?"

[Petronski]

Hermann the Cherusker developed this 15-day moving average based on derived raw data from Rasmussen.

[Plutarch]

In my opinion, the polls were being deliberately gimmicked, in hopes of helping Kerry.

This would require collusion between competing organizations. Why go through periods of helping Kerry, and then ultimately let the truth slip?

Conspiracies are ropes of sand. There may be a general bias amongst the pollsters, but they are unlikely to be colluding.

[StJacques]

I think there is a problem with the polls, but I tend to believe that the pollsters are not as much to blame as those who hire them. There are two issues that must be examined about polls; what "weighting" methodology may be used, if any, and when are the polls taken. The "weighting" issue was something I raised earlier in a critique of Zogby's methodology, which you can "pop up" and read here:

http://www.freerepublic.com/focus/f-news/1242103/posts

"Weighting" poll results can be a key factor for those who hire pollsters. If you want to inflate Kerry's numbers, all it takes is a weighting proportion that overestimates the turnout among Democrats - or those who identify themselves as Democrats - to get it done. For this reason you need to examine the underlying methodology used in a poll before judging it, and this may account for the variations between polls when you compare "weighted" and "non-weighted" results.

A second issue, and this applies to polls that are not weighted by party identification, is "when was the poll taken" and this one concerns me right now more than any other, because I'm wondering if CNN and USA Today are muzzling Gallup and preventing them from showing the true results. The last Gallup poll taken showed Kerry leading Bush by one point, but this poll was taken on Oct. 9 - 10, which was a Saturday and Sunday, two days when results tend to favor Democrats, as has been pointed out repeatedly within other threads on this site and can be found in a quick web search of other sites. Why have we not seen a Gallup poll on the national race since this one? Have CNN and USA Today told Gallup that they want to release up-to-date poll results on the national race on Mondays, which will mean that Gallup will again do their polling over the weekend and continue to favor Kerry in their results? Why did Gallup not release new state polling information this past Friday, as they have done every Friday for the past several weeks? Have CNN and USA Today buckled under to MoveOn.org for their public criticism of the role the news organizations are playing in using Gallup's polls without weighting the results by party identification? And I also want to point out that the recent polls of Midwestern states done for the Chicago Tribune which returned results quite favorable to Kerry were also primarily done over the weekend.

We must start paying attention to these matters. It is getting serious.

[infidel29]

What Kerry needing help to win this election? The hell you say.

[rwfromkansas]

One of the news values of journalism is conflict. A close race has that fighting slugmatch feel that journalists love because it breeds excitement in the newsroom.

It may not be to help Kerry at the core, but that is a side benefit to a liberal journalist.

[rwfromkansas]

That said, I do not believe in some massive pollster conspiracy.

However, I read the entire article and it does not claim that since the changes occurred at expected times.

[bobsunshine]

Thanks for the graphic. Using a 15 or 10 day moving average is the only way to really represent the data. Please continue to update. I think with Kerry's remark about Cheney's daughter and Edwards' comments about "walking again" with stem cell research, the numbers for Mr. Bush should increase as these remarks get talked about again on Sunday News shows.

[Freesofar]

What else could be the motive, the "base turnout" motive would be driven more by fear of losing, so if close then more turnout ?

Also believable would be the motive to make it look close and start another Florida style election controversy.

But all these conspiracy theories have a common problem,too many eyes and ears to spill the beans.

I'll buy that media parallax view factor, of pick and choose statistics reporting

After all "If it ain't close they can't cheat" (Hugh Hewitt) and like a runaway football game, there's an exodus for the turnstiles

[over3Owithabrain]

Sorry, the polls suck right now because Bush tanked the first debate. He snatched a possible defeat from the jaws of victory with that dreadful appearance. He has recovered some, but we won't know for sure until 11/2.

[Petronski]

I actually learned how to do graphs on excel just for this election season. I prepare the graph and get it on the web.


But the concept and calculations are that of Hermann the Cherusker, and bless him for doing it. That graph says it all: in the macro sense, Kerry has not been 'in this' since the RNC in NYC, thanks in part to his electoral meltdown at the hands of the soldiers of truth: the Swift Boat Vets.

[Read2Know]

"There may be a general bias amongst the pollsters, but they are unlikely to be colluding"
You can substitute the words "news media" for "pollsters" and come to the same conclusion.
The bias of the MSM is blatantly obvious in spite of no observable collusion. When you put together teams of people with similar beliefs, the group-think factor kicks in.

[Petronski]

We have in economics the concept of 'tacit collusion.' Example: airlines cut prices all at once. There's no meeting, but if one cuts prices and none follow, the price cuts end.

No one suggests a secret meeting among the alphabet network news bigs to plan the polls. But they do rub elbows at gatherings, and more importantly, they all know what they want done. Thus, Dan Rather can falsify 60 Minutes reports to slander the President and Doug Halperin can instruct his minions to go easier on lurch...they both can do this without some kind of secret lodge meeting to coordinate their efforts.

[Hermann the Cherusker]

This would require collusion between competing organizations. Why go through periods of helping Kerry, and then ultimately let the truth slip?

What I find interesting about the polling trends is the contrast between the "steady as she goes" types (Rasmussen, Fox/Opinion Dynamics, Zogby, IBD/Tipp, ARG) and the roller-coaster polls (Gallup/CNN/USA Today, CBS, Newsweek, Time, ABC/Washington Post, AP).

It seems to me a clear case of the professional pollsters on one side, and the liberal media and their darling Gallup on the other.

I don't say this meaning that some of the professionals are not biased (like Zogby and ARG), but that there are two totally different views of the race being pushed out there, one of which is definitely news-cycle driven.

The roller-coaster view has 15-20% of the country wildly swinging in its opinion day-by-day. Do any of us believe that 1 in 5 or 1 in 6 Americans has wildly changed their opinion on this race several times (early August, late August, late September, etc.).

More to the point, cumulative state-by-state polling by all outfits has not been swinging wildly, but has shown the race varying between one or the other candidate being up 3-5 points at most, which is the most the professional Pollsters and Fox have ever shown Bush or Kerry up.

Is it a conspiracy? No, I don't think so. But do the news groups have a common ideological agenda which might push them all to act in the same direction? Yes, that seems obvious. The biggest confirmation of this is probably the fact that the national conservative media - FOX, IBD (with TIPP) and the Wall St. Journal (with Zogby) are not onboard with the liberal media in their view of the race but are in bed with the professional Pollsters, which is where you would expect people of integrity to be.

As a conservative, I'm going to stick with the view taken by the conservative media outlets, as I have here on FR since August, which is confirmed by the state-by-state polling. In the end, this group will be shown far more accurate in their overall portrayal.

And you don't need to get sick on the liberal media roller-coaster along the way.

[Doctor Stochastic]

...the liberal media and their darling Gallup ...

These people belive that the Gallup Poll is run by Right Wing Fundamentalist Christians in order to derail the Kerry campaign.

[Hermann the Cherusker]

Thanks for the graphic. Using a 15 or 10 day moving average is the only way to really represent the data. Please continue to update.

Here is the Rasmussen data for a 15 and 6 day moving average of the race (roughly 2 week and 1 week cycles).

Hopefully, our good Mr. Petronski can graph this update for us. The 15 day numbers are first, and the 6 day numbers are second.

Date	Bush	Kerry	Bush	Kerry
2-Jun	44.80	45.00	45.00	44.00
3-Jun	44.60	44.80	44.50	44.00
4-Jun	45.00	44.80	45.00	44.50
5-Jun	44.80	45.00	44.50	45.00
6-Jun	44.80	44.80	45.00	44.50
7-Jun	45.00	44.80	44.50	45.00
8-Jun	44.80	45.00	45.00	45.00
9-Jun	44.80	44.80	45.00	45.50
10-Jun	45.00	44.80	45.00	45.00
11-Jun	45.40	44.40	46.00	44.50
12-Jun	45.20	44.60	45.50	45.00
13-Jun	45.40	44.80	46.00	45.00
14-Jun	45.40	45.20	46.50	45.50
15-Jun	45.40	45.00	46.50	45.00
16-Jun	45.40	45.00	46.50	45.00
17-Jun	45.80	45.20	46.00	46.00
18-Jun	45.80	45.20	46.00	45.50
19-Jun	45.60	45.20	45.50	45.50
20-Jun	46.00	45.40	46.00	45.50
21-Jun	45.40	45.60	45.00	46.00
22-Jun	45.40	45.60	44.50	46.50
23-Jun	45.80	45.80	45.00	46.50
24-Jun	46.00	45.40	45.50	46.00
25-Jun	45.60	45.60	45.00	46.00
26-Jun	45.60	46.20	45.00	46.50
27-Jun	45.80	45.80	46.50	45.50
28-Jun	45.20	45.60	45.50	45.00
29-Jun	45.40	45.80	45.50	45.50
30-Jun	45.60	45.80	45.50	46.00
1-Jul	45.00	45.80	45.00	45.50
2-Jul	45.40	46.00	45.50	45.50
6-Jul	45.80	46.00	46.00	46.00
7-Jul	45.40	45.60	46.00	45.50
8-Jul	45.40	46.00	46.00	46.00
9-Jul	46.20	46.00	46.50	46.50
10-Jul	45.60	46.00	46.00	47.00
11-Jul	45.60	46.20	45.50	47.00
12-Jul	45.60	46.60	45.00	47.50
13-Jul	45.40	46.60	45.00	48.50
14-Jul	45.40	46.60	45.00	48.00
15-Jul	45.40	47.00	44.50	48.00
16-Jul	45.60	47.20	45.50	48.00
17-Jul	45.60	47.00	45.50	47.50
18-Jul	45.60	47.00	45.50	47.00
19-Jul	46.00	47.00	46.50	46.50
20-Jul	45.80	46.80	46.50	46.00
21-Jul	45.60	47.00	46.50	46.00
22-Jul	45.60	47.60	46.00	46.50
23-Jul	45.60	47.20	46.00	46.50
24-Jul	45.40	47.00	46.00	46.50
25-Jul	45.80	47.20	45.50	47.50
26-Jul	45.80	46.80	45.50	47.00
27-Jul	45.80	46.80	45.50	47.00
28-Jul	45.80	47.20	45.50	47.50
29-Jul	45.80	46.80	45.50	47.00
30-Jul	45.80	46.80	45.50	47.50
31-Jul	45.80	47.00	45.50	47.50
1-Aug	45.60	47.20	45.00	48.50
2-Aug	45.80	47.00	45.50	47.50
3-Aug	45.40	47.60	45.50	47.50
4-Aug	45.40	47.40	45.50	47.50
5-Aug	45.40	47.40	45.50	47.50
6-Aug	45.60	47.60	45.50	48.00
7-Aug	45.80	47.20	46.50	46.50
8-Aug	46.00	47.40	46.50	47.50
9-Aug	46.00	47.60	47.00	47.50
10-Aug	45.80	47.80	46.50	48.00
11-Aug	46.00	47.80	47.00	48.00
12-Aug	46.00	47.60	46.50	47.50
13-Aug	46.00	47.80	46.00	48.50
14-Aug	46.20	47.60	46.00	48.00
15-Aug	46.00	47.80	45.50	48.00
16-Aug	46.40	47.60	46.50	48.00
17-Aug	46.20	48.00	46.00	48.00
18-Aug	46.20	47.80	46.00	48.00
19-Aug	46.40	48.00	46.50	48.00
20-Aug	46.60	48.00	46.50	48.50
21-Aug	46.40	47.60	46.50	47.50
22-Aug	46.40	48.00	46.50	47.50
23-Aug	46.20	48.00	46.50	47.50
24-Aug	46.20	47.60	47.00	47.00
25-Aug	46.60	47.60	47.00	47.00
26-Aug	46.40	47.40	46.50	46.50
27-Aug	46.40	47.20	46.50	46.50
28-Aug	46.80	47.20	47.00	46.50
29-Aug	46.80	47.00	47.50	45.50
30-Aug	46.60	46.80	46.50	46.00
31-Aug	46.80	46.80	47.00	46.00
1-Sep	47.00	46.60	47.50	46.00
2-Sep	47.20	46.20	48.00	45.50
3-Sep	47.40	46.20	48.00	45.50
4-Sep	47.42	45.94	48.05	45.85
5-Sep	47.32	46.08	48.30	45.70
6-Sep	47.52	46.10	48.30	45.75
7-Sep	47.68	46.00	48.20	46.00
8-Sep	47.56	45.98	47.90	46.45
9-Sep	47.62	46.06	47.55	46.65
10-Sep	47.84	46.04	47.55	46.75
11-Sep	47.86	46.00	47.85	46.30
12-Sep	47.88	45.90	47.90	46.00
13-Sep	47.76	46.32	47.70	46.30
14-Sep	47.88	46.10	47.30	46.30
15-Sep	47.94	45.98	47.80	45.80
16-Sep	48.22	45.86	48.45	45.55
17-Sep	47.92	46.14	48.15	45.85
18-Sep	48.02	45.94	48.35	45.60
19-Sep	47.96	46.14	48.55	45.40
20-Sep	48.02	45.92	48.65	45.25
21-Sep	48.16	45.74	48.85	45.15
22-Sep	48.26	45.64	48.30	45.45
23-Sep	48.18	45.70	48.55	45.35
24-Sep	48.14	45.68	47.85	46.00
25-Sep	48.28	45.66	48.35	45.55
26-Sep	48.12	45.78	48.10	45.95
27-Sep	48.04	45.86	47.60	46.30
28-Sep	48.34	45.64	47.90	46.30
29-Sep	48.52	45.54	48.15	45.90
30-Sep	48.28	45.80	48.15	46.10
1-Oct	48.22	45.76	48.30	45.80
2-Oct	48.40	45.62	48.85	45.45
3-Oct	48.20	45.92	48.75	45.75
4-Oct	48.38	45.76	48.65	45.70
5-Oct	48.36	45.96	48.25	46.30
6-Oct	47.98	46.20	48.10	46.15
7-Oct	48.18	46.14	48.20	46.40
8-Oct	48.24	46.10	48.15	46.55
9-Oct	48.42	46.08	48.40	46.40
10-Oct	48.50	45.98	48.65	46.10
11-Oct	48.60	45.88	48.70	45.75
12-Oct	48.34	46.02	48.50	45.85
13-Oct	48.44	45.96	48.55	45.85
14-Oct	48.48	46.00	48.75	45.65
15-Oct	48.44	45.90	48.20	45.65
16-Oct	48.36	46.14	47.95	46.20

The 15 day average sample shows the overall trend, while the 6 day average shows a clear sampling sine-wave.

The one and only inflection point of any signifance in Rasmussen's data occurred in the period August 23-27 when the slight advantage Kerry had been holding all summer suddenly vanished in a tidal wave (relatively speaking) of Bush support.

What happened then? National exposure of the Swift-Vets ad and story, and Kerry's pathetic response to it.

Frankly, had Kerry come out and said something like: "The Vietnam War was a difficult time for all of us. I respect my fellow veteran's views and recollections that differ from my own, although I disagree with them. At this difficult time for our nation, we need to focus on what will unite us in the future, not what divded us in the past when many mistakes were made by good and honest people on both sides who felt they were doing what was right for the country. I apologize if my actions at that time hurt any of my fellow veterans or fellow Americans. I, like they, was only trying to do what I felt was best for the country we all love."

Had Kerry done that, he'd be the one leading in the polls right now, and he'd be looking magnanimous, while the Swift-Vets would be looking like creeps.

Of course he couldn't because he's "John Kerry, reporting for duty."

Stick a fork in him. Short of a new DUI-type revelation in the next two weeks, come 11/3, he's back to being the junior Senator from Massachusetts. There are no major events out there which can change the dynamics of the race at this point, short of a scandal revelation against Bush, or Kerry performing a Christ-like miracle on live TV.

Bush has the support of 48-50% of the population, and at least 1/3 of the undecideds. Kerry has 45-46% and no more than 2/3 of the undecideds. The Field has 1-2%. You can't win when over 50% of the country won't be voting for you. And Kerry isn't the one with over 50%, since he is stuck at the same 45-46% level he had back in June and July.